无损压缩 + 质量对比与体积统计工作台
目标：把 PNG/JPEG/WebP 做尽量小的体积优化，同时给出可解释的对比结果，避免“看不出差异但体积更大”的反向优化。
英文 Prompt:
You are a senior full-stack engineer. Build a production-ready web app called "Image Optimizer Workbench".
Core features:
- Upload multiple images (png/jpg/jpeg/webp). Show list with name, dimensions, original size.
- Optimize/compress images with presets (lossless for PNG when possible; quality slider for JPEG/WebP).
- Before/after preview (side-by-side) + ive stats: bytes saved, percent saved.
- Export optimized files as a ZIP.
- Provide a "regression guard": detect if optimized file is larger than original and mark it; allow keep-original.
Tech constraints:
- Use Next.js (App Router) + Type .
- Image processing should run server-side with sharp; include safe limits for file size and pixel count.
- Add rate limiting and basic security headers.
Deliverables (must include all):
1) Full project file tree.
2) Complete source code for every file.
3) Local run commands (pnpm recommended).
4) Dockerfile + docker-compose.yml.
5) Deployment guide for a basic VPS and for Vercel (if serverless constraints apply, explain alternatives).
6) Tests: at least 5 automated tests OR a QA checklist of 10 items covering edge cases.
Do NOT include any image generation or AI art. Only process existing uploaded images.
中文释义: 让 AI 直接生成一个可部署的图片无损/有损压缩工作台，支持批量上传、对比预览、体积统计、打包下载，并带测试或 QA 清单。
EXIF 隐私清理 + 字段审计与可追溯报告工具
目标：一键去除照片定位、设备型号等隐私字段，并输出清理前后字段差异报告，方便团队合规留档。
英文 Prompt:
Build a web tool "EXIF Privacy Cleaner".
Features:
- Upload one or many JPEG/HEIC/PNG (if data exists) and parse data.
- Show a table of EXIF/IPTC/XMP fields and highlight risky fields (GPS*, serial, device model).
- Provide actions: "Strip all data" and "Strip only risky fields".
- After processing, allow download cleaned files and export a JSON audit report:
{ originalFile, removedFields, keptFields, timestamp, checksumBefore, checksumAfter }.
Implementation requirements:
- Node.js + Express API + Type .
- Use a reliable data library (exiftool-vendored or equivalent). Avoid shelling out to system exiftool.
- Add server-side validation and size limits.
Deliverables:
- Full repository code + file tree.
- Run commands + env vars.
- Docker build/run.
- QA checklist (>=10) including: files without EXIF, corrupted EXIF, large images, unicode filenames.
中文释义: 生成一个能审计并清理 EXIF 隐私字段的在线工具，既能下载清理后的文件，也能导出审计报告用于合规记录。
批量转格式：PNG/JPEG/WebP/AVIF 互转与透明通道规则校验
目标：做一站式格式转换并明确透明通道、颜色配置文件等规则，避免“转换后发灰/透明丢失”。
英文 Prompt:
Create "Image Format Converter" web app.
Requirements:
- Batch convert between PNG, JPEG, WebP, and AVIF.
- For each conversion, show warnings:
- Transparency will be lost when converting to JPEG.
- Color profile handling (force sRGB option).
- Provide presets and per-file overrides.
- Output ZIP.
Tech:
- Remix or Next.js + Type .
- Processing with sharp.
Deliverables:
- File tree + full code.
- Local run commands.
- Docker + deployment notes.
- Tests or QA checklist (>=10) focusing on transparency, alpha, color profiles, very small and very large images.
No image generation. Only convert existing images.
中文释义: 让 AI 产出一个批量图片格式转换工具，带透明通道与色彩规则提示、ZIP 下载、可部署与可验证。
批量尺寸调整：等比缩放、裁剪策略与“二次压缩”风险提示
目标：提供可重复、可解释的缩放/裁剪策略，并提示会触发二次压缩的场景，减少质量事故。
英文 Prompt:
Build "Batch Resize & Crop" tool.
Features:
- Upload many images.
- Choose mode: contain/cover/fill, target width/height, background color for padding.
- Optional: keep original data or strip.
- Display a dry-run plan per file (original dims -> output dims).
- Warn when resizing JPEG may cause recompression artifacts; allow quality control.
Stack:
- Next.js + Type .
- sharp server-side.
Deliverables:
- Full code + file tree.
- Commands + Docker.
- QA checklist (>=10) including panoramas, square crops, tiny icons, orientation EXIF.
中文释义: 生成一个可部署的批量缩放/裁剪在线工具，能显示每个文件的尺寸变更计划并控制质量。
颜色一致性：ICC 配置检测与强制 sRGB 转换工具
目标：检测并修正非 sRGB 的图片，降低跨设备显示偏色风险，输出修正前后信息对比。
英文 Prompt:
Build "Color Profile Inspector".
Features:
- Upload images and detect embedded ICC profile and color space.
- One-click convert to sRGB (optional) and strip ICC (optional) with clear warnings.
- Show before/after data summary.
Tech:
- Node + Type API.
- sharp or similar that supports ICC transforms.
Deliverables: full repo code, file tree, run commands, Docker, and QA checklist (>=10).
中文释义: 生成一个检测/修正图片色彩配置文件的在线工具，解决偏色与兼容性问题，支持输出对比信息与下载。
重复图片检测：感知哈希（pHash）聚类与误判复核工作台
目标：在不上传到第三方的前提下，对一批图片做相似度聚类，帮助清理重复素材。
英文 Prompt:
Create "Duplicate Image Finder".
Features:
- Upload a batch of images.
- Compute perceptual hash (pHash/dHash) and group similar images by threshold.
- Provide UI to review groups and mark which files to keep/delete (client-side selection), then export a CSV report and ZIP of selected kept images.
Implementation:
- Prefer client-side hashing using WebAssembly or pure JS to reduce server load; if server-side, include safeguards.
- Next.js + Type .
Deliverables: complete code + file tree + run/deploy instructions + QA checklist with at least 10 cases.
中文释义: 生成一个基于感知哈希的重复图片检测与复核工具，支持相似度阈值、聚类展示、导出报告与打包下载。
图片差异对比：像素级 Diff 热力图与阈值标注工具
目标：用于回归测试与设计验收，快速看到两张图的差异区域，并导出差异报告。
英文 Prompt:
Build "Image Diff Reporter".
Features:
- Upload two images ( line and current) or upload two folders with matching filenames.
- Generate pixel diff with threshold and produce a heatmap overlay.
- Output: diff images + a JSON report summarizing changed pixels and mismatched files.
Tech:
- Node + Type .
- Use pixelmatch (or equivalent) for diff.
- Provide CLI-friendly API endpoint too.
Deliverables: full source + file tree, run commands, Docker, and at least 5 automated tests (preferred) or a QA checklist (>=10).
中文释义: 生成一个图片差异对比工具，支持单对比或按文件名批量对比，输出热力图和结构化报告，适合回归测试。
水印添加：批量文字/二维码水印、位置策略与可撤销配置导出
目标：给素材或对外分享图加水印（不做去水印），支持可复现的参数配置与批量处理。
英文 Prompt:
Create "Batch Watermark Adder" (adding watermarks only; never remove).
Features:
- Batch upload images.
- Add text watermark (font, size, opacity, rotation) and optional QR code watermark.
- Placement presets (top-left, bottom-right, tiled) with safe margins.
- Export a watermark-config JSON so results are reproducible.
Stack:
- Next.js + Type + sharp.
Deliverables: full repo code + file tree + run commands + Docker + QA checklist (>=10).
中文释义: 生成一个批量加水印的在线工具，支持文字与二维码水印、位置策略和参数配置导出，明确“不做去水印”。
图片体检：尺寸/格式/透明/元数据/文件名规则批量校验器
目标：用于团队交付前检查，批量发现不合规图片（过大、格式不对、含隐私元数据、命名不规范）。
英文 Prompt:
Build "Image QA Validator".
Features:
- Upload a ZIP of images or a batch of files.
- Configure rules: max bytes, max dimensions, allowed formats, forbid data, filename regex.
- Output a validation report (HTML + JSON) listing pass/fail with reasons.
- Provide a one-click remediation suggestion list (e.g., convert to WebP, strip EXIF).
Tech: Node + Type + Next.js.
Deliverables:
- Full code + file tree.
- Run/deploy commands + Docker.
- Automated tests (>=5) OR QA checklist (>=10) including malformed filenames and mixed encodings.
中文释义: 生成一个图片批量质检工具，支持规则配置、报告导出与整改建议，适合素材交付与上线前检查。
